IPv6:: Internet 2

IPv4, on which the internet is based, is flawed. Within a decade its address space will be exhausted, and the Internet will cease to grow. The solution is IPv6, and you can access it here and now.

When the net was first conceived it was originally intended as an academic and military research network. It was, then, a project of the US Department of Defence Advanced Research Projects Agency (ARPA). However by the time the first implementation began in 1969, almost seven years since the concept was first suggested by computer scientist J.C.R Licklider at MIT in 1962 (in which he described a 'galactic network'), it became clear that the fledgling project could expand to include a much wider network outside the original brief.

It was on this ARPA network (later known as ARPANET three years later that the first email in the history of the world was sent in 1971. In 1973, file transfers followed with the development of FTP, and the building blocks of the internet were founded. There was even a specification for a Network Voice Protocol (NVP), what we know today as VoIP, but wouldn't become technically feasible for another three decades.

The network ran using the original protocol suite that was desgined to carry information between machines. Termed the 'Network Control Program (NCP)', it put the load for communication on servers and formed the basis for computers to communicate using what was simply called the 'host-to-host protocol'.

It was quickly realised that as the network grew the NCP servers would have increasingly larger loads placed upon them, and that NCP wasn't going to cut it. As early as 1973 the specification for a new protocol suite was drafted by the now famous Vint Cerf and Bob Kahn that would place the load on every client and not on servers. The key to it was the Transport Control Protocol (TCP), designed to allow software processes to communicate with eachother over a network. It first saw light in 1977 and gradually replaced NCP. In 1978 the Internet Protocol (IP), responsible for addressing and routing packets between disparate machines, was added to TCP and the combination we now know as TCP/IP became the dominant protocol on ARPANET.

Then in January 1983, nine years since the building blocks were first laid down in Vint Cerf's paper, TCP/IP is adopted as the only protocol for ARPANET. At the same time the Department of Defence's network, known as MILNET, is split off from ARPANET for security. All the while the network continued to grow at a rapid pace as more and more machines in academic institutions from around the world joined up.

And the rest, as they say, is history. It would continue to expand and eventually in 1991, after years of work, Tim Berners-Lee revealed the fledgling pieces of software that would enable the World Wide Web -- the first Web browser, the first Web server and of course the first Web pages, which described the project itself. By the mid 90s commercial interests would join the wave leading to the mass adoption of the internet and WWW as a global information network.

The need for bigger boots
Today, the world is the net. It's become not just a facet of society as we know it, but a vital block upon which it is built. And as it continues to influence every part of our lives, it's only going to get bigger.

Which brings us to the problem: the current specification of TCP/IP, known as IPv4, just doesn't allow for a network the size of which we are rapidly approaching. There are more people and computers in the world than there are addresses to assign to them. As TCP/IP is a point to point protocol, every machine on the net needs a unique address. If there are no spare addresses, a machine simply can't join it.

And IPv4 is wasteful. Addresses when registered are assigned in classes, for which a 'Class C' address proivdes 256 unqiue IPs (XXX.XXX.XXX.255) while a 'Class B' provides 65536 IPs (XXX.XXX.255.255). If you need 300 for your company, Class C isn't enough and Class B is a vast waste of the address space.

The advent of NAT (Network Address Translation) has, naturally, eased this. NAT allows local networks to conntect through a single machine on the net, and therey use only one address instead of many. If you're running a LAN at home sharing a connection to the Internet through a broadband modem or router, you're using NAT.

But while NAT is alleviating the shortfall on address space it's actually considered a Bad Thing -- NAT introduces a single point of failure for a whole network of machines as well as breaking true end-to-end connections between computers, the very purpose of TCP/IP to begin with.

Finally IPv4 is inherently insecure, providing no means of encryption or security at the protocol layer. Some would argue this is better handled at the application layer regardless, but it will certainly provide a wider scope for protection in a world increasingly moving commerce and economy online through interconnected private networks.

In other words while IPv4 launched the net, it's served its purpose and is no longer suited to the task that lies ahead - an every growing, ever dependable, global network to power the world.